Evaluate¶

Evaluation functions for model performance metrics.

evaluate¶

openddi.evaluate.evaluate.plot_metrics(train_metrics, val_metrics, metric_name, out_file_prefix)[source]: Plot training and validation metrics

_metrics_from_logits¶

openddi.evaluate.evaluate._metrics_from_logits(y_true_np, y_logits_np)[source]¶: Return Acc/F1/Recall/Precision (macro average, avoiding undefined warnings)

_metrics_from_logits_multilabel¶

openddi.evaluate.evaluate._metrics_from_logits_multilabel(y_true_np, y_logits_np)[source]¶: Return AUC and AP (macro average, avoiding undefined warnings)

plot_metrics¶

openddi.evaluate.evaluate.plot_metrics(train_metrics, val_metrics, metric_name, out_file_prefix)[source]¶: Plot training and validation metrics